Rebalancing volumes in computer system comprising upper and lower cluster

Abstract

In a computer system including a distributed file system cluster and a block SDS cluster, the distributed file system cluster has a plurality of distributed file system nodes, and stores a management-subject file redundantly in a plurality of volumes managed by a plurality of distributed file system nodes, the block SDS cluster has a plurality of block SDS nodes, and provides a plurality of the volumes on the basis of storage regions of storage apparatuses of the block SDS nodes, and a CPU of a management server is configured to perform control such that a plurality of the volumes storing the file in the distributed file system cluster for redundancy are not volumes based on a storage region of a storage apparatus of one block SDS node.

What is claimed is: 
     
       1. A computer system comprising an upper cluster and a lower cluster, wherein
 the upper cluster has a plurality of first nodes, and stores a management-subject data unit redundantly in a plurality of volumes managed by a plurality of first nodes, 
 the lower cluster has a plurality of second nodes, and provides a plurality of the volumes on a basis of storage regions of storage apparatuses of the second nodes, 
 a processor of the computer system performs control such that a plurality of the volumes storing the data unit in the upper cluster for redundancy are not volumes based on a storage region of a storage apparatus of one second node, 
 the processor of the computer system 
 identifies a second node having a storage region on which each of a plurality of the volumes storing the data unit for redundancy is based, and 
 in a case where one second node is identified for a plurality of the volumes, moves some volumes that are included in a plurality of the volumes and are based on a storage region of the one identified second node to a second node other than the one identified second node. 
 
     
     
       2. The computer system according to  claim 1 , wherein
 the processor of the computer system performs control such that all of a plurality of the volumes storing the data unit for redundancy are volumes based on storage regions of storage apparatuses of different second nodes. 
 
     
     
       3. The computer system according to  claim 2 , wherein
 the redundancy is attained by multiplication of the data unit. 
 
     
     
       4. The computer system according to  claim 2 , wherein
 the redundancy is attained by erasure coding. 
 
     
     
       5. The computer system according to  claim 1 , further comprising a managing apparatus, wherein
 a processor of the managing apparatus 
 identifies a second node having a storage region on which each of a plurality of the volumes storing the data unit for redundancy is based, and 
 in a case where one second node is identified for a plurality of volumes, instructs a processor of a second node in the lower cluster to move some volumes that are included in a plurality of the volumes and are based on a storage region of the one identified second node to a second node other than the one identified second node. 
 
     
     
       6. The computer system according to  claim 1 , further comprising a managing apparatus, wherein
 a processor of the managing apparatus transmits, to the lower cluster, redundancy volume information that can identify a plurality of the volumes storing the data unit for redundancy, and 
 a processor of a second node in the lower cluster 
 receives the redundancy volume information, 
 identifies a second node having a storage region on which each of a plurality of the volumes corresponding to the redundancy volume information is based, and 
 in a case where one second node is identified for a plurality of volumes, moves some volumes that are in a plurality of the volumes and are based on a storage region of the one identified second node to a second node other than the one identified second node. 
 
     
     
       7. The computer system according to  claim 2 , wherein
 the processor of the computer system 
 receives redundancy volume plurality of the volumes storing the data unit for redundancy, and 
 generates a plurality of volumes corresponding to the redundancy volume information such that the generated plurality of volumes are volumes based on storage regions of storage apparatuses of different second nodes. 
 
     
     
       8. A volume arrangement managing apparatus that manages movement of volumes in a computer system including an upper cluster and a lower cluster, wherein
 the upper cluster has a plurality of first nodes, and stores a management-subject data unit redundantly in a plurality of volumes managed by a plurality of first nodes, 
 the lower cluster has a plurality of second nodes, and provides a plurality of the volumes on a basis of storage regions of storage apparatuses of the second nodes, 
 a processor of the managing apparatus performs control such that a plurality of the volumes storing the data unit in the upper cluster for redundancy are not volumes based on a storage region of a storage apparatus of one second node, 
 the processor of managing apparatus 
 identifies a second node having a storage region on which each of a plurality of the volumes storing the data unit for redundancy is based, and 
 in a case where one second node is identified for a plurality of the volumes, moves some volumes that are included in a plurality of the volumes and are based on a storage region of the one identified second node to a second node other than the one identified second node. 
 
     
     
       9. A volume arrangement management method performed by a computer system including an upper cluster and a lower cluster, the method comprising:
 storing, by the upper cluster having a plurality of first nodes, a management-subject data unit redundantly in a plurality of volumes managed by a plurality of first nodes; 
 providing, by the lower cluster having a plurality of second nodes, a plurality of the volumes on a basis of storage regions of storage apparatuses of the second nodes; 
 performing control, by the computer system, such that a plurality of the volumes storing the data unit in the upper cluster for redundancy are not volumes based on a storage region of a storage apparatus of one second node; 
 identifying by the computer system a second node having a storage region on which each of a plurality of the volumes storing the data unit for redundancy is based, and 
 in a case where one second node is identified for a plurality of the volumes, moving by the computer system some volumes that are included in a plurality of the volumes and are based on a storage region of the one identified second node to a second node other than the one identified second node.